Web Services Description Teleconference
18 Mar 2004


See also: IRC log


   David Booth            W3C
   Allen Brookes          Rogue Wave
   Roberto Chinnici       Sun Microsystems
   Glen Daniels           Sonic Software
   Paul Downey            British Telecommunications
   Youenn Fablet          Canon
   Yaron Goland           BEA Systems
   Hugo Haas              W3C
   Tom Jordahl            Macromedia
   Jacek Kopecky          Systinet
   Amelia Lewis           TIBCO
   Jonathan Marsh         Chair (Microsoft)
   Ingo Melzer            DaimlerChrysler
   Jean-Jacques Moreau    Canon
   Bijan Parsia           University of Maryland MIND Lab
   Igor Sedukhin          Computer Associates 
   Asir Vedamuthu         webMethods
   Sanjiva Weerawarana    IBM
   Umit Yalcinalp         Oracle
   Prasad Yendluri        webMethods, Inc.
   Arthur Ryman           IBM
   William Vambenepe      Hewlett-Packard
   Jerry Thrasher         Lexmark
Jonathan Marsh
Paul Downey



<JacekK> Jonathan, I'll have to leave in about an hour

Approval of Minutes

<Scribe> minutes from joint meeting with XMLP WG approved

<Scribe> minutes from Cannes F2F approved

<Scribe> minutes from 11th March telecon approved

Action Item Review


  PENDING   2003-09-18: Marsh to review the QA operational

  PENDING   2004-01-08: Pauld to write up examples of schemas for the

  PENDING   2004-01-28: Philippe and JMarsh will look at the ipr for 
                        test suite.

  PENDING   2004-01-30: DaveO to write up a proposal for augmenting 
                        schema information to enable versioned data.

  PENDING   2004-02-12: DaveO to produce a refined proposal for Asynch 
                        HTTP binding addressing the concerns of folks 
                        that object to leaving replyTo info out of WSDL.

  DONE [.2] 2004-02-26: Editors to adopting the diff in Gudge's proposal
                        along with the two additional words to 
                        Section 3.2. (#143) 

  DONE [.3] 2004-02-26: Editors to rename the message property to 
                        "element" (#142).

  DONE [.4] 2004-02-26: Editors to clarify wording (#122).

  EDTODO    2004-03-04: Editors to add back the WSDL Component 
                        Designators back to the spec as SHOULD.

  DONE [.5] 2004-03-04: Editors to incorporate the above proposal to 
                        issue 79.

  DONE[.17]  2004-03-04: DBooth to propose specific changes needed for 
                        processor conformance text.

  DONE [.6] 2004-03-04: Primer Editors to include best practices and the 
                        Schema group to review the best practices 

  DONE[.15] 2004-03-05: Editors to clarify the spec to say that
                        required attribute means that a feature 
                        must be understood and it must be engaged.

  DONE[.15] 2004-03-05: Editors to clarify that the strongest value of
                        the required attribute wins.

  DONE [.7] 2004-03-05: Editors to add clarification text with regards to 
                        issue 121.

  EDTODO    2004-03-05: Editors to write in the spec the wsdlLocation
                        global attribute proposal along what has been 
                        done for xsi:schemaLocation.

  DONE [.8] 2004-03-05: Editor to add optional Boolean "safe" attribute to
                        interface operation, corresponding property in 
                        the component model.

  PENDING   2004-03-05: DavidO to notify TAG about our decision on safety.

  DONE [.8] 2004-03-05: Editors to further clarify the spec to make clear
                        that all operations must be bound.

  DONE [.9] 2004-03-05: Editors to clarify the spec to say that its ok to
                        have bad imports but if during QName resolution 
                        you can't find something then it fails like any 
                        other qname resolution issue.

  DONE[.14] 2004-03-05: Editors to update draft to say that include will
                        fail early.

  DONE[.14] 2004-03-05: Editors to add text saying two imports with same 
                        import/@targetNamespace is legal.

  DONE [.9] 2004-03-05: Editors to clarify that if an optional extension 
                        (i.e., an extension not marked as required) is 
                        not understood it MUST be ignored. Any not 
                        understood extension attributes MUST be ignored.

  DONE      2004-03-05: Editors (namely Gudge and Roberto) to update the 
                        schema and the spec accordingly.

  DONE[.14] 2004-03-05: Editors to change spec to require absolute URIs
                        and indicate that comparison must be done 
                        character-by-character as per TAG finding.

  DONE      2004-03-05: Jonathan to send a note to XMLP, asking them for 
                        volunteers for editors.

  PENDING   2004-03-05: Editors of media type proposal to give Jonathan a 
                        list of open issues.

  PENDING   2004-03-11: Sanjiva to fill in data for May FTF on logistics 

  DONE[.10] 2004-03-11: Editors to clarify 2.4.1 on why no global qname 
                        reference to endpoints.

  DONE[.11] 2004-03-11: JMarsh to find part of spec impacted by 115 and send 
                        mail out to group.

  PENDING   2004-03-11: JMarsh, David Booth, David Orchard to form adhoc 
                        group to explore stylistic rendering options.

  DONE[.14] 2004-03-11: Editors to change @name to @ref as found in email 

  DONE[.14] 2004-03-11: Editors to change @label to @messagelabel as 
                        proposed by Sanjiva.

  DONE[.12] 2004-03-11: Editors to add a {value} property to the Property 

  DONE[.13] 2004-03-11: JMarsh to provide component model changes supporting 
                        #any|#empty proposal.

  DONE[.16] 2004-03-11: Jacek to propose resolution for issue 152,
                        Importing the targetNamespace.

  [.1] http://www.w3.org/2002/ws/desc/#actions
  [.2] http://lists.w3.org/Archives/Public/www-ws-desc/2004Mar/0095.html
  [.3] http://lists.w3.org/Archives/Public/www-ws-desc/2004Mar/0096.html
  [.4] http://lists.w3.org/Archives/Public/www-ws-desc/2004Mar/0097.html
  [.5] http://lists.w3.org/Archives/Public/www-ws-desc/2004Mar/0100.html
  [.6] http://lists.w3.org/Archives/Public/www-ws-desc/2004Mar/0105.html
  [.7] http://lists.w3.org/Archives/Public/www-ws-desc/2004Mar/0102.html
  [.8] http://lists.w3.org/Archives/Public/www-ws-desc/2004Mar/0104.html
  [.9] http://lists.w3.org/Archives/Public/www-ws-desc/2004Mar/0144.html
  [.10] http://lists.w3.org/Archives/Public/www-ws-desc/2004Mar/0094.html
  [.11] http://lists.w3.org/Archives/Public/www-ws-desc/2004Mar/0106.html
  [.12] http://lists.w3.org/Archives/Member/w3c-ws-desc/2004Mar/0052.html
  [.13] http://lists.w3.org/Archives/Public/www-ws-desc/2004Mar/0073.html
  [.14] http://lists.w3.org/Archives/Public/www-ws-desc/2004Mar/0165.html
  [.15] http://lists.w3.org/Archives/Public/www-ws-desc/2004Mar/0171.html
  [.16] http://lists.w3.org/Archives/Public/www-ws-desc/2004Mar/0117.html
  [.17] htp://lists.w3.org/Archives/Public/www-ws-desc/2004Mar/0064.html

New Issues - Import/Include

Yaron: WSDL depends upon retrieval of data. unhygenic methods of loading this data. publisher of a popular WSDL must be rich! proposal supports load balancing

Marsh: wonders how href works at all

Yaron: slashdot effect: if you're popular - you're dead!

Jacek: Web depends upon data retrieval - this is bigger than WSDL

Amy: + to Jacek. even parsers now have catalogue properties - including open source products widely distributed

<umit> +1 to Amy

Tom: +1 to Jacek and Amy

Sanjiva: nothing in our spec to restrict load-balancing ATM

<sanjiva> I also think that just allowing a list of URIs doesn't *really* solve this problem properly: all you've gone is gone from a list of length 1 to a another *finite* list. If the list is finite every problem you had before still exists.

Yaron: further explains load balancing and redundancy as posted to list - nasty performance for those at the top of the list of locations - DNS round-robin doesn't work well thanks to caching

<sanjiva> I'd like to request we do a straw poll and go on .. this is a small issue to spend so much time on.

Yaron: forces big-iron on those publishing popular WSDL

<asir> +1 to strawpoll

<GlenD> +1. There are lots of other small issues we could spend so much time on. :)

Jacek: suggests URI to actual location mapping possible outside of WSDL

<umit> I request a straw poll to see whether this is a problem we should solve

<sanjiva> I'm absolutely against making a new issue for this! We discussed this at the F2F and decided not to do it and I don't see why it should be dealt with now.

<sanjiva> (or in the future)

Marsh: do we want to track this as an issue: 1) return to issues list for later discussion or 2) drop this now

<GlenD> Let's see what the Last Call folks think.

<umit> +1 to Sanjiva

<GlenD> I am STRONGLY hoping that the soapbuilders community re-ignites into a flurry of good discussion and activity once our LC drafts come out.

<GlenD> Issues like this should be discussed there amongst the implementers as well as here amongst the standards-wonks. :)

<sanjiva> +1 for SOAPBuilders and others really giving this stuff a run thru and generating good last call comments.

New Issues - RPC signature tokens from ## to #

<sanjiva> +1 to s/##/#/g

<Roberto> imagine the number of keystrokes you save!

Moving normative text from notes to separate paragraph

<sanjiva> +1!

dbooth: +1

<sanjiva> ACTION: editors to implement David Booth's proposal on making all notes non-normative and moving the normative text into the regular paragraphs.

New Issues - messageLabel attribute optional in interface, but not in binding

<Roberto> just do it!

Marsh: accepted Sanjiva's proposal

<sanjiva> ACTION: editors to make binding/operation/(input|output)/@messageLabel optional and have the same rules as the corresponding thing in interface/operation for computing the value.

Issue 112 New headers/body style proposal

Tom: this seems like overkill

<sanjiva> +1 for being overkill / hard to understand

Marsh: should possibly be baked into WSDL

<asir> +1 to more time

Sanjiva: wants more time to consider this

<GlenD> baking it into WSDL is certainly another option

Tom: use me as a litmus test for our target audience

Umit: no relationship on how this integrates into properties

Glen: correct: this is about how application data maps into headers - it's not a generic for spec writers

Umit: requests clear text for distinction

<TomJordahl> More specifically, I read the proposal twice and didn't really understand the mechanism and uses for it. This is a red flag as far as how our 'customers' might react to this complexity

<prasad> Add me to the list of don't understand. Need more time to digest..

<sanjiva> Sanjiva: it should be possible to indicate that there's more data flowing by using a new MEP with multiple input messages and then defining a binding for that which puts some of the messages in headers and some on the body.

Roberto: requests more time, worries about the levels of nesting, the validation requirements and the complexity therin

Marsh: sounds like we're not ready to deal with this - leave it on our issues list

Yaron: being forced to go down the F&P route introduced the complexity, wants his requirement addressed

<sanjiva> We can take a vote if necessary and see what the WG thinks ... I certainly will not vote for adopting this solution.

<Roberto> ditto, it needs improvement

Marsh: this is a controversial issue

Tom: we did go through this before Yaron joined the WG

Marsh: we need to move on to part 3 to meet our schedule.

Yaron: happy to work with whoever to progress this issue

Marsh: issue remains open

Issue 152 - importing the targetNamespace

Jacek: proposes to close issue with no action

dbooth: suggest non-normative best practice text

<Scribe> http://lists.w3.org/Archives/Public/www-ws-desc/2004Mar/0117.html

Tom: prefers to take action for a processor not to have to enforce the prohibition

Sanjiva: enforcement is a string comparison

otherwise we would have to state rules for how duplicates from the same namespace should be handled

Jacek: would like the 'class of processors' to fail on this

Marsh: seems like we need more complexity to allow this

dbooth: blanket statement processor must fault if encounters this

discussion on change in semantics regarding optional extensions

<TomJordahl> http://www.w3.org/2002/ws/desc/2/06/issues.html#x115

<sanjiva> (probably)

Marsh: sounds like we have concensus

<Scribe> ACTION: editors to clarify the meaning of 'change' WRT to optional extensions

<dbooth> Proposed rewording for issue 115 sentence:

<dbooth> [[

<dbooth> The presence of an optional extensibility element or attribute MAY therefore augment the semantics of a WSDL document in ways that do not invalidate the existing semantics. However, the presence of a mandatory extensibility element MAY alter the semantics of a WSDL document in ways that invalidate the existing semantics.

<dbooth> Note: Authors of extensibility elements SHOULD avoid altering the existing semantics in ways that are likely to confuse users.

<dbooth> ]]

Issues 146 and 150 - operations with *anything* and nothing in the body

Asir: what is the value of message content model if other schema models used

Marsh: parallel for other schema languages but component model allows different schema models to do whatever they want

element attribute a bit confusing if element can point to nothing

Sanjiva: suggests using extensibility model to introduce better syntax

<asir> arthur's e-mail is at http://lists.w3.org/Archives/Public/www-ws-desc/2004Mar/0113.html

Marsh: can't think of better name than "element" at this point, good enough for this draft, but could consider other names later

Tom: suggests element="none"

<asir> yeh, element="#none"

Robert: types system other than XML schema, always need to choose one of these three values, other type systems may not have an "element"

Roberto: content model property has no value indication you're not using XMLSchema, explicity allow property to have no value

Umit: our definition could be broken if we allowed this. suggests we should do this through extensibility understands use-case for empty messages, but "any" ?

<yaron> +1

Marsh: logging service, can send them whatever, header describes routing, but body could contain "any"

Umit: grave concerns there would be no schema to describe message

Sanjiva: element="none" makes sense, element="any" can be handled by extensibility

<prasad> +1 to Sanjiva's suggestion

<umit> +1 to Sanjivae

Yaron: where is the limit of extensibility ? we have a clear use-case of an intermediary, logging, time-stamp, encryption, etc.

<pdowney> +1 to Yaron!

<sanjiva> How about element=qname for a single element, element=#none, and body=#any?

Umit: grave concerns - this breaks all the good we have done!

Tom: speaks in support of this proposal, good use-case

<Scribe> straw poll: Jonathan's proposal element= none, qname or any

Marsh: 7 yes, 1 no, several abstains

Umit: objects

<sanjiva> ACTION: editors to incorporate Jonathan's proposal for issues 146 & 150 (with #empty changed to #none)

Publication of working drafts for parts 1&2 ?

jjm: should be able to get something soon

<Scribe> group agrees to approve publication of a working drafts

Sanjiva: why discuss HTTP binding ahead of SOAP binding ?

Marsh: been more traffic on HTTP binding

jjm: has been looking at schema for SOAP binding

Summary of Action Items

[NEW] ACTION: editors to clarify the meaning of 'change' WRT to optional
[NEW] ACTION: editors to implement David Booth's proposal on making all
  notes non-normative and moving the normative text into the regular
[NEW] ACTION: editors to incorporate Jonathan's proposal for issues 146 &
  150 (with #empty changed to #none)
[NEW] ACTION: editors to make binding/operation/(input|output)/@messageLabel
  optional and have the same rules as the corresponding thing in
  interface/operation for computing the value.

Minutes formatted by David Booth's scribe.perl 1.63 (CVS log)
$Date: 2004/03/16 18:33:33 $